Skip to main content

Full text of "Iron Emission in the z=6.4 Quasar SDSS J114816.64+525150.3"

See other formats


Preprint typeset using liTgX style emulateapj v. 25/04/01 



IRON EMISSION IN THE Z = 6.4 QUASAR SDSS Jl 14816.64+525150.3 

Aaron J. Barth'-^, Paul Martini', Charles H. Nelson*, and Luis C. Ho' 



m 
o 
o 



ABSTRACT 

We present near-infrared J and /T-band spectra of the z = 6.4 quasar SDSS Jl 14816.64+525150.3 obtained 
with the NIRSPEC spectrograph at the Keck-II telescope, covering the rest-frame spectral regions surrounding 
the C IV A1549 and Mg II A2800 emission lines. The iron emission blend at rest wavelength 2900-3000 A is 
clearly detected and its strength appears nearly indistinguishable from that of typical quasars at lower redshifts. 
The Fe II/Mg II ratio is also similar to values found for lower-redshift quasars, demonstrating that there is no 
strong evolution in Fe/a broad-line emission ratios even out to z = 6.4. In the context of current models for iron 
enrichment from Type la supernovae, this implies that the SN la progenitor stars formed at z > 10. We apply 
the scaling relations of Vestergaard and of McLure & Jarvis to estimate the black hole mass from the widths of 
the C IV and Mg II emission lines and the ultraviolet continuum luminosity. The derived mass is in the range 
(2-6) X lO' Mq, with an additional uncertainty of a factor of 3 due to the intrinsic scatter in the scaling relations. 
This result is in agreement with the previous mass estimate of 3 x 10^ Mq by Willott, McLure, & Jarvis, and 
supports their conclusion that the quasar is radiating close to its Eddington luminosity. 

Subject headings: quasars: emission Unes — quasars: individual (SDSS Jl 14816.64+525150.3) 



> 

in 
o 
o 

00 
O 
m 
o 

■ 

6 



1. INTRODUCTION 

The broad emission lines of high-redshift quasars are lu- 
minous beacons that can be used to study the metal enrich- 
ment history of the densest regions in the early universe (e.g., 
Hamann & Ferland 1999). Observations of emission lines 
such as C IV and N V reveal supersolar abundances in quasar 
broad-line regions (BLRs) even at z > 4 (Dietrich et al. 2003a), 
strengthening the link between the growth and fueling of black 
holes and the buildup of stellar mass in the host galaxies. The 
abundance ratio of iron to a elements is a key diagnostic that 
may be used as a "clock" to constrain the earliest epoch of star 
formation (Hamann & Ferland 1993; Yoshii et al. 1998), since 
enrichment of a elements occurs via core-collapse supernovae 
while the dominant channel for iron enrichment is Type la su- 
pernovae, which have longer-lived progenitor stars. 

The flux ratio of the integrated iron emission in the 2200- 
3000 A wavelength range to Mg II AA2796,2802 is an obser- 
vationally accessible diagnostic that is dependent on the un- 
derlying Fe/Mg abundance ratio, but also sensitive to the BLR 
density and microturbulence (Verner et al. 2003). Numerous 
studies have demonstrated that there is no systematic decrease 
in the Fe II/Mg II flux ratio of quasars with increasing redshift, 
even out to z w 6 (e.g., Thompson, Hill, & Elston 1999; Diet- 
rich et al. 2002, 2003b; Freudling et al. 2003), while Iwamuro 
et al. (2002) found evidence for an increase in Fe II/Mg II with 
increasing redshift to z ~ 5. 

SDSS Jl 14816.64+525150.3 (abbreviated here as SDSS 
Jl 148+5251) was discovered by Fan et al. (2003) in the Sloan 
Digital Sky Survey; at z = 6.4 it is the highest-redshift quasar 
currently known. Willott et al. (2003) used H and K-hwA spec- 
tra of this object to obtain an estimate of the black hole mass 
(~ 3 X 10^ Mq) from the Mg II linewidth, using the scaling rela- 
tions derived by McLure & Jarvis (2002). The Mg II A2800 line 
was the only emission feature clearly present in their spectrum; 

' Caltech Optical Observatories, 105-24 Caltech, Pasadena, CA 91125 
2 Hubble Fellow 

^ The Observatories of the Carnegie Institution of Washington, 813 Santa Barbara Street, Pasadena, CA 91101 
■* Department of Physics and Astronomy, Drake University, 2507 University Avenue, Des Moines, lA 5031 1-4505 

1 



the limited S/N of their data did not permit a definitive detection 
of the surrounding Fe II emission blends. The quasar environ- 
ment was evidently a site of vigorous massive star formation 
and metal enrichment prior to z = 6.4, as demonstrated by the 
roughly normal equivalent width of the Mg II emission line as 
well as the presence of submillimeter emission from 10** Mq 
of dust in the quasar host galaxy (Bertoldi et al. 2003). Here, 
we present new near-infrared spectra of SDSS Jl 148+5251 ob- 
tained at the Keck Observatory, covering the rest-frame spectral 
regions surrounding the C IV A 1549 and Mg II A2800 emission 
lines. We show that its Fe II/Mg II ratio is similar to that of 
typical quasars at lower redshifts, and we update the black hole 
mass estimate of Willott et al. using the higher-quality Keck 
data. Except where noted otherwise, we assume Hq = 70 km s~' 
Mpc"\ VL„, = 0.3, and 51a = 0.7; for this cosmology the universe 
at z = 6.4 was 840 Myr old. 

2. OBSERVATIONS AND REDUCTIONS 

The observations were obtained using the NIRSPEC spec- 
trograph (McLean et al. 2000) at the Keck-Il telescope on the 
nights of 2003 March 1 1 and 12 UT. We used a Of'76-wide slit 
in the NIRSPEC- 1, 2, and 6 settings, which cover the wave- 
length ranges 0.95-1.12, 1.08-1.29, and 1.89-2.31 Aim, respec- 
tively, at /? « 1500. A standard ABBA nodding sequence was 
used with exposure times of 300 s at each nod position. The 
airmass ranged from 1.2 to 1.5. Seeing, as measured from the 
spatial profiles of calibration star exposures, was typically be- 
tween 0"7 and 0"9. The total exposure times in each setting 
were 6000, 9600, and 8400 s, respectively. Due to hardware 
problems, the nodding and guiding performance of NIRSPEC 
was relatively poor during this run and slit losses were severe, 
possibly ^ 50% during some individual exposures. 

The spectra were extracted from bias-subtracted, flattened 2- 
dimensional frames using an optimal weighting algorithm, and 
using the sky-subtraction code of Kelson (2003). Wavelength 



2 



BARTH ET AL. 



calibration was performed with the OH airglow emission lines 
in each exposure. The data were flux-calibrated and corrected 
for telluric absorption using spectra of the AOV star HD 99966, 
observed immediately before or after the quasar in each setting, 
following the methods described by Vacca, Gushing, & Rayner 
(2003). Finally, the spectra were scaled to the J and K' magni- 
tudes given for SDSS Jl 148H-525 1 by Fan et al. (2003), to place 
them on an absolute flux scale. 

3. RESULTS AND DISCUSSION 

The spectra are displayed in Figure 1, overplotted with the 
SDSS composite quasar spectrum of Vanden Berk et al. (2001). 
The best match between the Mg II profile of SDSS Jl 148H-525 1 
and the SDSS composite is found for a redshift of 6.40 ± 0.01, 
in agreement with the previous measurement of z = 6.41 ± 0.01 
by Willott et al. (2003). However, the C IV emission line ap- 
pears strongly blueshifted with respect to the Mg II redshift and 
is much broader than the C IV line of the composite quasar. 
The C IV line falls within a telluric absorption band at 1.14 /xm, 
making accurate flux calibration difficult, but the extended blue 
wing of C IV does appear to be a real feature in the spectrum. 
In quasars, Mg II is considered a good indicator of the systemic 
redshift while relative C IV blueshifts of 500-2000 km s"' or 
more are conmion (Marziani et al. 1996; Richards et al. 2002). 
The centroid of the C rv line measured by a Gaussian fit corre- 
sponds to a blueshift of 2900 ±120 km s"' relative to the Mg II 
line. Richards et al. (2002) find that the Si IV A 1400 emission 
line of quasars is typically not strongly blueshifted, unlike C IV; 
Figure 1 shows that SDSS J 1148-1-5251 appears to follow this 
trend. 

The data do not reveal any broad absorption lines, but a few 
narrow absorption lines are present. The strongest are at 1.0853 
/Ltm (equivalentwidth 5.6±0.8 A) and at 1.1842 fjm (equivalent 
width 4.4 ± 0.4 A). Neither is coincident with an atmospheric 
absorption or emission feature, and they appear to be genuine 
features in the quasar spectrum. The line at 1.1842 /xm is unre- 
solved, while the 1.0853 /xm feature is marginaUy broader than 
the night sky emission lines but not broad enough to be consis- 
tent with a blended C IV or Mg II doublet. The identification of 
these fines is uncertain, and neither of them matches any wave- 
lengths that would be expected for metal absorption lines from 
the z = 4.943 C IV absorber found by White et al. (2003). 

3.1. Fe II Emission 

In general, measurement of the iron emission strength is best 
accomplished when the data have wide spectral coverage, so 
that spectral regions without either iron emission or Balmer 
continuum emission can be included in the fit (e.g., Dietrich 
et al. 2002). Our spectrum does not extend to sufficiently long 
wavelengths to do this, and we fit only the rest-frame spectral 
region 2610-3100 A. An Fe II emission blend at 2900-3000 A 
is clearly present in the Keck spectrum. From the comparison 
between the NIRSPEC data and the SDSS composite spectrum 
in Figure 1, it is already apparent that the Fe n/Mg II ratio in 
SDSS 1148-H5251 is roughly similar to that in the composite 
quasar. 

The spectrum was modeled as the sum of three components: 
a power-law continuum, an empirical iron emission template, 
and a double-Gaussian model for the Mg II profile. For the 
Mg II model, aU of the Gaussian parameters were allowed to 

^ The Vestergaard & Wilkes (2001) iron template is not distributed or shared freely, 
article. 



float freely. We do not ascribe any particular physical inter- 
pretation to the two components in terms of the structure of 
the emission-line regions; this is simply the minimal empiri- 
cal description that allows an adequate fit to the emission line. 
Dietrich et al. (2002) show that the Mg II profiles of quasars 
can generally be fit well by a Gaussian near the systemic ve- 
locity plus another blueshifted (and usuaUy broader) Gaussian 
component. We note that the model fit may be affected by an 
imperfect correction for telluric absorption in the blue wing of 
Mg II. Given our limited spectral coverage we chose not to in- 
clude the Balmer continuum emission as a separate component, 
since its strength could not be determined independently of the 
dominant power-law continuum. 

The iron blends in this wavelength region are traditionally 
modeled using an empirical "template" spectrum derived from 
an AGN with intrinsically narrow emission lines. We used 
an iron template derived from HST ultraviolet observations of 
I Zw 1 as described by Vestergaard & Wilkes (2001)^; the iron 
emission is dominated by Fe II blends with a contribution from 
Fe III. The iron template was broadened by convolution with 
a Gaussian having the same width as the stronger Mg n com- 
ponent. To optimize the fit we computed a parameter de- 
termined using the la uncertainties on each pixel in the flux- 
calibrated spectra. Figure 2 displays the best-fitting model. The 
power-law index of the model continuum is a = -1.4 ± 0.1 for 
fx oc A", similar to the value of a = -1.56 for the SDSS com- 
posite quasar (Vanden Berk et al. 2001). 

Following Dietrich et al. (2002), the Fe II/Mg II emission-line 
ratio was determined by integrating the iron template flux over 
the wavelength range 2200-3090 A. Although this requires ex- 
trapolation of the scaled iron template beyond the wavelength 
range over which our fit was performed, it allows the most 
straightforward comparison with previous measurements. We 
find Fe II/Mg II = 4.7 ± 0.4. This result is similar to values that 
have been found for quasars over all observed redshifts. For 
example, Dietrich et al. (2002) obtained Fe II/Mg II = 3.3^.2 
for six quasars at z « 3 .4 and Fe II/Mg II = 3 .8 ± 0.4 for a low- 
z composite quasar, while Thompson, Hill, & Elston (1999) 
measured Fe II/Mg II « 6 - 7 (over a slightly wider wavelength 
range of 2000-3000 A for Fe II) for composite quasar spectra 
at z = 3 and z = 4, and 4.3 ± 0.1 for a low-z composite quasar. 
Our result is also within the same range as values measured by 
Iwamuro et al. (2002), Dietrich et al. (2003b), and Freudling et 
al. (2003) for quasars at z = 4.4-6.3. 

Thus, the lack of any strong evolution in Fe II/Mg II noted 
by Thompson et al. out to z = 3 - 4 apparently continues out 
to z = 6.4; even at this redshift there is still no indication of a 
substantial decrease in this ratio. If the Fe II/Mg II flux ratio 
can be interpreted as even a rough indicator of the underlying 
abundance ratio, then the age of the universe at z = 6.4 provides 
some constraints on metal enrichment models. The minimum 
Ufetime for SN la progenitors has often been considered to be 
~ 1-1.5 Gyr; this would imply that the Fe/a abundance ratio 
should undergo a strong rise starting at 1-1.5 Gyr following 
the first burst of star formation in which SN la progenitors are 
created (Hamann & Ferland 1993; Yoshii et al. 1998). 

Observations of Fe II/Mg II ratios in quasars at z ~ 4 could 
still be consistent with such models (Dietrich et al. 2002), but 
our observations and those of FreudUng et al. (2003) rule out 



so we re-created their template spectrum using information given in their published 



IRON EMISSION IN SDSS J114816.64+525150.3 



3 



models with iron enrichment timescales of > 800 Myr. How- 
ever, the timescale for the maximum SN la rate (and hence iron 
enrichment) depends sensitively on the star formation rate and 
initial mass function, and may be much shorter than the canoni- 
cal 1 Gyr for an intense starburst in the core of a rapidly forming 
elhptical galaxy (Fria9a & Terlevich 1998; Matteucci & Recchi 

2001) . The models of Friafa & Terlevich (1998) give a typ- 
ical timescale of 0.3 Gyr to reach solar abundance of iron in 
the interstellar medium of a proto-elliptical galaxy; similarly, 
Matteucci & Recchi (2001) find that the peak SN la rate in an 
elhptical galaxy occurs ~ 0.3 Gyr after the major burst of star 
formation. If the iron abundance were built up in just 0.3 Gyr, 
then the SN la progenitors would have formed at z > 10. There 
may be only a brief temporal window during which a quasar 
would exhibit a low Fe/a abundance ratio, and this epoch could 
be at a redshift as high as ~ 10 if star formation began at z « 20 
(e.g., Kogut et al. 2003). Alternatively, the iron may have 
been produced in pair-instability supernovae from Population 
111 stars with initial masses of 140-260 Mq; a single such ex- 
plosion could produce up to 40 Mq of iron (Heger & Woosley 

2002) . This would result in iron enrichment of the pregalactic 
gas within a few Myr after the first burst of star formation. 

3.2. The Mass of the Black Hole 

The only practical methods to estimate the black hole masses 
in distant quasars are based on scaling relations derived from 
low-redshift samples. Reverberation mapping of Seyfert galax- 
ies and quasars at z < 0.4 has demonstrated a correlation be- 
tween continuum luminosity and BLR radius (Wandel, Peter- 
son, & Malkan 1999; Kaspi et al. 2000). By combining the 
continuum luminosity with measurements of broad-hne widths, 
the black hole mass can be estimated under the assumption of 
virial motion of the BLR clouds. The original scaling relations 
were derived from FWHM(H;3) and the continuum luminos- 
ity at 5 100 A, but similar relations have been determined using 
ultraviolet continuum luminosity combined with the Unewidth 
of either C IV (Vestergaard 2002) or Mg II (McLure & Jarvis 
2002). These relations have the form M, oc [XLx{\5V)]'^ x 
FWHM(line)2, where (3 = 0.47 (McLure & Jarvis 2002) or 0.7 
(Vestergaard 2002). The dominant uncertainty in the M, es- 
timates comes from the intrinsic scatter in the M,-linewidth 
correlations, which is a factor of ^ 2.5-3, rather than from 
measurement uncertainties on the linewidths. 

Willott et al. (2003) found FWHM(Mg II) = 6000!jjg° km s"' 
based on a double-Gaussian fit with the two components fixed 
at rest wavelengths of 2796 and 2802 A, giving M, « 3 x 10^ 
Mq. Our spectrum shows that the profile is rather asymmet- 
ric, and the full Mg II profile has FWHM(Mg II) = 5500 ± 200 
km s~^ Applying the McLure & Jarvis (2002) relation with 
ALa(3000 A) = 5.7 X 10^'' W, we obtain M. w 2 x 10*^ Mq. This 
result is close to the previous estimate by Willott et al., since we 
find similar values for FWHM(Mg II) and La(3000 A). 

The peculiar shape of the C IV hue might result from non- 
Keplerian kinematics, calling into question the application of 
the Vestergaard (2002) method to derive M, for this particu- 
lar object. Keeping this caveat in mind, we will derive an M, 
estimate from the C rv Une for purposes of comparison with 
the Mg II method. To measure FWHM(C IV), the spectrum was 
first prepared by subtracting off the Fe template, with its overall 
scaling determined by the fit to the /T-band spectrum. The Fe 
emission only makes a small (~ 10%) contribution to the flux 
level in the continuum surrounding C IV. Then, the region from 



1450 to 1630 A rest wavelength was fit with a single Gaus- 
sian plus continuum. The best-fit Gaussian has FWHM(C IV) = 
9000 ± 300 km s"'. Vestergaard's M, relation was derived as- 
suming a cosmology with Ho = 75 km s~^ Mpc~', qo = 0.5, and 
A = 0. Taking ALa(1350 A) = 2.5 x 10'*'' erg s"' for these cos- 
mological parameters, we obtain M, x6x 10^ Mq for SDSS 
J1148-H5251. Given that these scahng methods have an esti- 
mated uncertainty of a factor of 2.5-3, the two estimates of M, 
from Mg II and C IV are in reasonable agreement. It would still 
be worthwhile to obtain a C IV spectrum with higher S/N ratio 
to check the reality of the large blueshift and linewidth. 

To estimate the bolometric luminosity, we apply a bolometric 
correction of L\,oi/\iyL,^(2500 A)] = 6.3, which is typical of low- 
redshift quasars (Elvis et al. 1994). This gives Lboi « 4 x 10"*^ 
erg s"' and implies (as previously shown by Willott et al.) an 
Eddington ratio of Lboi/iEdd « 1 forM, = 3 x 10^ Mq. 

A few additional caveats should be noted: the apparent lu- 
minosity of the quasar can be affected by dust extinction, con- 
tinuum beaming, or gravitational lensing. Significant extinc- 
tion is unhkely because the spectral shape of the quasar does 
not appear strongly reddened relative to the SDSS compos- 
ite quasar, while Willott et al. (2003) argue against continuum 
beaming because the Mg II equivalent width is typical of un- 
beamed quasars. Lensing is expected to affect only a small 
fraction of high-z SDSS quasars (Wyithe & Loeb 2002; Pin- 
dor et al. 2003), and infrared images of SDSS J1148H-5251 do 
not give any indication of lensing (Fan et al. 2003). 

It is remarkable to find such a massive black hole at such a 
large look-back time. Nevertheless, models for the hierarchical 
growth of supermassive black holes by merging and gas accre- 
tion can accommodate masses of > 10^ Mq in quasars at z > 6 
if the seed black holes form at z > 10 (e.g., Haiman & Loeb 
2001 ; Volonteri, Haardt, & Madau 2003). The initial seed black 
holes may be objects with M, « 100-200 Mq formed by col- 
lapse of supermassive Population III stars (Fryer, Woosley, & 
Heger 2001; Madau & Rees 2001; Schneider et al. 2002). We 
note that even a single seed black hole having 180 Mq at z = 20 
could (just barely) grow to 3 x 10^ Mq by z = 6.4 if it were 
constantly accreting at the Eddington luminosity with a radia- 
tive efficiency of e = 0.1, so the present data do not yet require 
multiple black hole mergers to build up the observed black hole 
mass. 



4. CONCLUSIONS 

The Keck spectra reveal an Fe II/Mg II ratio in SDSS 
J 1 148H-525 1 that is within the normal range for quasars at lower 
redshifts, indicating that substantial iron enrichment must have 
occurred in the densest regions of the universe at redshifts 
greater than 6.4. Recent photoionization calculations by Verner 
et al. (2003) have shown that Fe II/Mg II is not a straightforward 
diagnostic of Fe/a abundance ratios, but useful constraints on 
the Fe/Mg abundance ratio can be derived from observations of 
the ultraviolet Fe n/Mg n ratio combined with the ratio of ul- 
traviolet to optical Fe II strength. The rest-frame optical Fe II 
emission in very high-redshift quasars cannot be measured from 
the ground, but these measurements can be pursued in the future 
with the James Webb Space Telescope. Searches for quasars at 
still higher redshifts may provide a glimpse of the earliest epoch 
of iron production and give stringent constraints on the forma- 
tion redshift of the seed black holes that were the progenitors 
of luminous quasars. 



4 



BARTH ET AL. 



We thank D. Kelson for the use of his sky-subtraction code 
and D. Stern for helpful comments. Research by A.J.B. is 
supported by NASA through Hubble Fellowship grant #HST- 
HF-01134.01-A awarded by STScI. RM. was supported by a 
Carnegie Starr Fellowship. Data presented herein were ob- 
tained at the W.M. Keck Observatory, which is operated as a 



scientific partnership among Caltech, the University of Cali- 
fornia, and NASA. The Observatory was made possible by the 
generous financial support of the W.M. Keck Foundation. The 
authors wish to recognize and acknowledge the very significant 
cultural role and reverence that the summit of Mauna Kea has 
always had within the indigenous Hawaiian community. 



REFERENCES 



Bertoldi, R, Carilli, C. L., Cox, P., Fan, X., Strauss, M. A., Beelen, A., Omont, 

A., & Zylka, R. 2003, A&A, in press 
Dietrich, M., Appenzeller, I., Vestergaard, M., & Wagner, S. J. 2002, ApJ, 564, 

581 

Dietrich, M., Hamann, F, Shields, J. C, Constantin, A., Heidt, J., Jager, K., 

Vestergaard, M., & Wagner, S. J. 2003a, ApJ, 589, 722 
Dietrich, M., Hamann, F, Appenzeller, I., & Vestergaard, M. 2003b, ApJ, in 

press 

Elvis, M., et al. 1994, ApJS, 95, 1 
Fan, X., et al. 2003, AJ, 125, 1649 

Freudling, W., Corbin, M. R., & Korista, K. T. 2003, ApJ, 587, L67 
Friafa, A. C. S. & Terlevich, R. J. 1998, MNRAS, 298, 399 
Fryer, C. L., Woosley, S. E., & Heger, A. 2001, ApJ, 550, 372 
Haiman, Z., & Loeb, A. 2001, ApJ, 552, 459 
Hamann, F, & Feriand, G. 1993, ApJ, 418, 11 
Hamann, F, & Feriand, G. 1999, ARA&A, 37, 487 
Heger, A. & Woosley, S. E. 2002, ApJ, 567, 532 

Iwamuro, F, Motohara, K., Maihara, T., Kimura, M., Yoshii, Y., & Doi, M. 
2002, ApJ, 565, 63 

Kaspi, S., Smith, P. S., Netzer, H., Maoz, D., Jannuzi, B. T., & Giveon, U. 2000, 

ApJ, 533,631 
Kelson, D. D. 2003, PASP, 1 15, 688 
Kogut, A., et al. 2003, ApJ, in press 
Madau, R, & Rees, M. J. 2001, ApJ, 551, L27 



Marziani, P., Sulentic, J. W., Dultzin-Hacyan, D., Calvani, M., & Moles, M. 

1996, ApJS, 104, 37 
Matteucci, F, & Recchi, S. 2001, ApJ, 558, 351 

McLean, I. S., Graham, J. R., Beckhn, E. E., Figer, D. F., Larkin, J. E., 

Levenson, N. A., & TepUtz, H. I. 2000, Proc. SPIE, 4008, 1048 
McLure, R. J., & Jarvis, M. J. 2002, MNRAS, 337, 109 
Pindor, B., Turner, E. L., Lupton, R. H., & Brinkmann, J. 2003, AJ, 125, 2325 
Richards, G. T., Vanden Berk, D. E., Reichard, T. A., Hall, R B., Schneider, D. 

R, SubbaRao, M., Thakar, A. R., & York, D. G. 2002, AJ, 124, 1 
Schneider, R., Rerrara, A., Natarajan, R, & Omukai, K. 2002, ApJ, 571, 30/ 
Thompson, K. L., Hill, G. J., & Elston, R. 1999, ApJ, 515, 487 
Vacca, W. D., Gushing, M. C, & Rayner, J. T. 2003, PASP, 115, 389 
Vanden Berk, D. E., et al. 2001, AJ, 122, 549 

Vemer, E., Bruhweiler, F, Vemer, D., Johansson, S., & GuU, T. 2003, ApJ, 592, 
L59 

Vestergaard, M. 2002, ApJ, 571, 733 

Vestergaard, M., & Wilkes, B. J. 2001, ApJS, 134, 1 

Volonteri, M., Haardt, F, & Madau, R 2003, ApJ, 582, 559 

Wandel, A., Peterson, B. M., & Malkan, M. A. 1999, ApJ, 526, 579 

White, R. L., Becker, R. H., Fan, X., & Strauss, M. A. 2003, AJ, 126, 1 

Willott, C. J., McLure, R. J., & Jarvis, M. J. 2003, ApJ, 587, L15 

Wyithe, J. S. B., & Loeb, A. 2002, ApJ, 581, 886 

Yoshii, Y., Tsujimoto, T., & Kawara, K. 1998, ApJ, 507, LI 13 



IRON EMISSION IN SDSS J114816.64+525150.3 



5 



3 F 



2.5 - 



1.5 - 



1 - 



T 1 1 r 



T IT 1 r 




Si IV 




1.1 



1.2 




-< 0.6 - 



2.1 



2.2 



2.3 



Observed Wavelength (yum) 



Fig. 1 . — Near-infrared spectra of SDSS Jl 148+5251. The upper panels show the combined NIRSPEC-1 and 2 settings, and the lower panels show the NIRSPEC-6 
setting. The spectrum overplotted on that of SDSS J1148+5251 is the SDSS composite quasar of Vanden Berk et al. (2001), transformed to z = 6.40. The same 
scaling factor has been applied to the SDSS composite quasar in both spectral regions. The panels below the quasar spectra show the atmospheric transmission and 
the night sky emission spectra. The Si IV, C IV, and Mg 11 einission hnes are labelled. There is iron emission throughout this spectral region; the prominent iron 
blend at 2900-3000 A rest wavelength is labelled. 



6 



BARTH ET AL. 



1 — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — r 




2600 2700 2800 2900 3000 3100 

Rest Wavelength (A) 

Fig. 2. — Model fit to the Mg II spectral region. The double-Gaussian model for the Mg II profile, the broadened Fe emission template, and the power-law 
continuum are displayed individually, and the full model is overplotted on the quasar spectrum. The observed spectrum has been shifted to the rest frame and the 
flux density has been scaled by (1 + z)'.